Whether the project is creating a new system from scratch, installing a new commercial software package, or making enhancements to an existing system, some level of information conversion and migration will probably need to occur. As a part of the design process, a clear strategy needs to be mapped out as to how this data conversion and system loading will be accomplished. Determining the approach consists of a three step process:
Before planning the appropriate conversion/migration, some form of a mapping effort should take place to cross reference the information required by the system to the anticipated sources of the data. These sources may include existing computerized databases, paper-based filing systems, personal computer spreadsheets, or just scraps of paper in people's desk drawers. In many cases, a part of the new system information may not need to be converted at all, it will simply be created from scratch. Along with determining these data sources, it is also useful to record each business area which has responsibility for the information.
The key to the mapping effort is to be specific whenever possible. Since the system is still being designed, either the proposed database file/table names or the related entity names make excellent reference points. Likewise, the anticipated data sources should be referenced in a similar fashion. The results of this mapping effort should be captured by utilizing either a cross-reference matrix or a migration diagram.
By looking ahead toward the programming and implementation of the system, a time-table should be developed which reflects the sequencing for the actual loading of the required information. The resulting time-table provides the foundation for the creation of a data loading workplan. Generally, this loading of the business information must be sequenced, since each wave of new data is usually validated against previously loaded information, and these dependencies must be taken into account. Additionally, if a phased implementation approach will be utilized, further impacts on the data loading sequencing may be realized.
Each individual source of information should be analyzed to determine the appropriate method for loading the data into the new system. Some of the potential techniques are listed below:
One-time data conversion programs
One-time, quick and dirty programs may need to be especially written to read information from the existing database files/tables in order to do any necessary data conversions and to load the data into the database files/ tables which are a part of the new system.
On-time data upload/download programs
Information which resides on a different technical platform than that of the new system may need to be transferred through an upload/download process. This may require the acquisition of special utility software or additional special programming.
Manual data entry
In some situations, special programming, or uploading/ downloading, is just not justified because either the quantity of the data to convert is very low, the technical problems are just too great, or the characteristics of the information have changed too much. This necessitates the manual entry of the information.